INDEX

Explanations

phrases related to widespread or notable events or trends

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "phenomenon" in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

gur

-0.91

bey

-0.75

zar

-0.66

ACTED

-0.65

gar

-0.64

bor

-0.64

ver

-0.62

 powdered

-0.61

zer

-0.61

min

-0.61

POSITIVE LOGITS

 phenomenon

1.28

 phenomena

1.05

ĸļ

0.98

ually

0.97

pmwiki

0.85

ptoms

0.80

interstitial

0.79

rences

0.78

DragonMagazine

0.76

 syndrome

0.76

Activations Density 0.007%

phrases related to widespread or notable events or trends

mentions of the term "phenomenon" in various contexts

No Comments

No Known Activations